the desire to limit family size.* Interestingly enough, Schultz finds in a predictive model for fertility in Taiwan that the association between birth rates and child school attendance rates is more consistent and statistically significant than between birth rates and the proportion of adults with a primary education (45).
Rising school attendance rates result, after a lag, in a rise in the level of education among the adult population, and this contributes to the decline in fertility in a number of ways, none of them completely independent of general socioeconomic development (46). On the personal level, educated couples are more likely to evaluate rationally the pros and cons of an extra birth, and perhaps be less concerned about the various taboos, cultural and religious, on the use of birth control.' They may also understand more clearly the conflicts between "quantity" and "quality" in the raising of children. These effects are likely to be reinforced if population and family life education has been in their school curriculum (47). Through their education they may have developed heightened aspirations and acquired new desires apart from child rearing, some of which directly conflict with child rearing. For example, mobility, which is restricted by a large family, is needed to realize fully the economic and "social status" benefits flowing from education. Whatever the motivation, there is no question that in the developing countries, better educated couples have fewer children and use contraception more than couples with little or no education.?
The economic burden is the most frequently cited disadvantage of a large family in surveys conducted in tropical Africa (43, Table 3). As part of this economic burden, the importance of educational costs "climbs steadily from subsistence to cash farming areas and from the latter to the towns, culminating among the urban elite where these costs are the major factor which could lead to the limitation of family size" (43, p. 604). As one specific example, Caldwell reports that in every part of rural Ghana a substantial majority of households reported in a 1963 survey that school attendance not only increases the cost of rearing children while decreasing their labor value, but that the upkeep of school children outweighs their productive value even in the case of 10-to-14-ycar-old day students. In the towns, the cost of education was sharply felt, particularly among the economically better off classes. Interestingly enough, part of the problem was that educated children tended to expect their parents to spend more on them, and the parents often felt obliged to meet the demand (44, and references cited therein).
t However, among Roman Catholics, extended education in parochial schools may serve to reinforce traditional Catholic values regarding fertility and contraception.
* Among surveys of knowledge, attitudes, and practice (KAP) of contraception either published or on file in the Population Council, all nineteen studies from eleven countries in Asia and Africa in which practice of contraception was cross-classified by educational attainment of wife, husband, or both showed a positive association between the two variables; in almost all cases the progression in percentages cither currently using or ever using contraception was an unbroken one from lowest to highest levels of educational attainment. The inverse relationship between education and fertility is not quite as clearcut, although of twenty-one studies in sixteen developing countries, most showed a steady downward progression in fertility as educational attainment increased, the impact of education on fertility tending to be more pronounced above the primary school level. In a few cases, the progression was broken between two levels of educational attainment;